10 - Deprecated ``admin_up`` field. Instead, ``admin_state`` is introduced,
11 with 3 possible values -- ``up``, ``down`` and ``offline``.
12 - Replaced ``--disks`` option of ``gnt-instance replace-disks`` with a
13 more flexible ``--disk`` option. Now disk size and mode can be changed
15 - Removed deprecated ``QueryLocks`` LUXI request. Use
16 ``Query(what=QR_LOCK, ...)`` instead.
17 - The LUXI requests :pyeval:`luxi.REQ_QUERY_JOBS`,
18 :pyeval:`luxi.REQ_QUERY_INSTANCES`, :pyeval:`luxi.REQ_QUERY_NODES`,
19 :pyeval:`luxi.REQ_QUERY_GROUPS`, :pyeval:`luxi.REQ_QUERY_EXPORTS` and
20 :pyeval:`luxi.REQ_QUERY_TAGS` are deprecated and will be removed in a
21 future version. :pyeval:`luxi.REQ_QUERY` should be used instead.
22 - CertificateError now derives from GanetiApiError in the RAPI client.
23 - Deprecation warnings due to pycrypto/paramiko import in
24 tools/setup-ssh have been silenced, as usually they are safe; please
25 make sure to run an up-to-date paramiko version
26 - The QA scripts now depend on Python 2.5 or above
32 *(Released Fri, 11 May 2012)*
34 A small bugfix release.
36 The main issues solved are on the topic of compatibility with newer LVM
39 - fixed parsing of ``lv_attr`` field
40 - adapted to new ``vgreduce --removemissing`` behaviour where sometimes
41 the ``--force`` flag is needed
43 Also on the topic of compatibility, ``tools/lvmstrap`` has been changed
44 to accept kernel 3.x too (was hardcoded to 2.6.*).
46 A regression present in 2.5.0 that broke handling (in the gnt-* scripts)
47 of hook results and that also made display of other errors suboptimal
48 was fixed; the code behaves now like 2.4 and earlier.
50 Another change in 2.5, the cleanup of the OS scripts environment, is too
51 aggressive: it removed even the ``PATH`` variable, which requires the OS
52 scripts to *always* need to export it. Since this is a bit too strict,
53 we now export a minimal PATH, the same that we export for hooks.
55 The fix for issue 201 (Preserve bridge MTU in KVM ifup script) was
56 integrated into this release.
58 Finally, a few other miscellaneous changes were done (no new features,
59 just small improvements):
61 - Fix ``gnt-group --help`` display
62 - Fix hardcoded Xen kernel path
63 - Fix grow-disk handling of invalid units
64 - Update synopsis for ``gnt-cluster repair-disk-sizes``
65 - Accept both PUT and POST in noded (makes future upgrade to 2.6 easier)
71 *(Released Thu, 12 Apr 2012)*
73 Incompatible/important changes and bugfixes
74 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
76 - The default of the ``/2/instances/[instance_name]/rename`` RAPI
77 resource's ``ip_check`` parameter changed from ``True`` to ``False``
78 to match the underlying LUXI interface.
79 - The ``/2/nodes/[node_name]/evacuate`` RAPI resource was changed to use
80 body parameters, see :doc:`RAPI documentation <rapi>`. The server does
81 not maintain backwards-compatibility as the underlying operation
82 changed in an incompatible way. The RAPI client can talk to old
83 servers, but it needs to be told so as the return value changed.
84 - When creating file-based instances via RAPI, the ``file_driver``
85 parameter no longer defaults to ``loop`` and must be specified.
86 - The deprecated ``bridge`` NIC parameter is no longer supported. Use
88 - Support for the undocumented and deprecated RAPI instance creation
89 request format version 0 has been dropped. Use version 1, supported
90 since Ganeti 2.1.3 and :doc:`documented <rapi>`, instead.
91 - Pyparsing 1.4.6 or above is required, see :doc:`installation
92 documentation <install>`.
93 - The "cluster-verify" hooks are now executed per group by the
94 ``OP_CLUSTER_VERIFY_GROUP`` opcode. This maintains the same behavior
95 if you just run ``gnt-cluster verify``, which generates one opcode per
97 - The environment as passed to the OS scripts is cleared, and thus no
98 environment variables defined in the node daemon's environment will be
99 inherited by the scripts.
100 - The :doc:`iallocator <iallocator>` mode ``multi-evacuate`` has been
102 - :doc:`New iallocator modes <design-multi-reloc>` have been added to
103 support operations involving multiple node groups.
104 - Offline nodes are ignored when failing over an instance.
105 - Support for KVM version 1.0, which changed the version reporting format
107 - TCP/IP ports used by DRBD disks are returned to a pool upon instance
109 - ``Makefile`` is now compatible with Automake 1.11.2
110 - Includes all bugfixes made in the 2.4 series
115 - The ganeti-htools project has been merged into the ganeti-core source
116 tree and will be built as part of Ganeti (see :doc:`install-quick`).
117 - Implemented support for :doc:`shared storage <design-shared-storage>`.
118 - Add support for disks larger than 2 TB in ``lvmstrap`` by supporting
119 GPT-style partition tables (requires `parted
120 <http://www.gnu.org/s/parted/>`_).
121 - Added support for floppy drive and 2nd CD-ROM drive in KVM hypervisor.
122 - Allowed adding tags on instance creation.
123 - Export instance tags to hooks (``INSTANCE_TAGS``, see :doc:`hooks`)
124 - Allow instances to be started in a paused state, enabling the user to
125 see the complete console output on boot using the console.
126 - Added new hypervisor flag to control default reboot behaviour
127 (``reboot_behavior``).
128 - Added support for KVM keymaps (hypervisor parameter ``keymap``).
129 - Improved out-of-band management support:
131 - Added ``gnt-node health`` command reporting the health status of
133 - Added ``gnt-node power`` command to manage power status of nodes.
134 - Added command for emergency power-off (EPO), ``gnt-cluster epo``.
136 - Instance migration can fall back to failover if instance is not
138 - Filters can be used when listing nodes, instances, groups and locks;
139 see *ganeti(7)* manpage.
140 - Added post-execution status as variables to :doc:`hooks <hooks>`
142 - Instance tags are exported/imported together with the instance.
143 - When given an explicit job ID, ``gnt-job info`` will work for archived
145 - Jobs can define dependencies on other jobs (not yet supported via
146 RAPI or command line, but used by internal commands and usable via
149 - Lock monitor (``gnt-debug locks``) shows jobs waiting for
152 - Instance failover is now available as a RAPI resource
153 (``/2/instances/[instance_name]/failover``).
154 - ``gnt-instance info`` defaults to static information if primary node
156 - Opcodes have a new ``comment`` attribute.
157 - Added basic SPICE support to KVM hypervisor.
158 - ``tools/ganeti-listrunner`` allows passing of arguments to executable.
160 Node group improvements
161 ~~~~~~~~~~~~~~~~~~~~~~~
163 - ``gnt-cluster verify`` has been modified to check groups separately,
164 thereby improving performance.
165 - Node group support has been added to ``gnt-cluster verify-disks``,
166 which now operates per node group.
167 - Watcher has been changed to work better with node groups.
169 - One process and state file per node group.
170 - Slow watcher in one group doesn't block other group's watcher.
172 - Added new command, ``gnt-group evacuate``, to move all instances in a
173 node group to other groups.
174 - Added ``gnt-instance change-group`` to move an instance to another
176 - ``gnt-cluster command`` and ``gnt-cluster copyfile`` now support
177 per-group operations.
178 - Node groups can be tagged.
179 - Some operations switch from an exclusive to a shared lock as soon as
181 - Instance's primary and secondary nodes' groups are now available as
182 query fields (``pnode.group``, ``pnode.group.uuid``, ``snodes.group``
183 and ``snodes.group.uuid``).
188 - Numerous updates to documentation and manpages.
190 - :doc:`RAPI <rapi>` documentation now has detailed parameter
192 - Some opcode/job results are now also documented, see :doc:`RAPI
195 - A lockset's internal lock is now also visible in lock monitor.
196 - Log messages from job queue workers now contain information about the
197 opcode they're processing.
198 - ``gnt-instance console`` no longer requires the instance lock.
199 - A short delay when waiting for job changes reduces the number of LUXI
200 requests significantly.
201 - DRBD metadata volumes are overwritten with zeros during disk creation.
202 - Out-of-band commands no longer acquire the cluster lock in exclusive
204 - ``devel/upload`` now uses correct permissions for directories.
210 *(Released Fri, 23 Mar 2012)*
212 This was the sixth release candidate of the 2.5 series.
218 *(Released Mon, 9 Jan 2012)*
220 This was the fifth release candidate of the 2.5 series.
226 *(Released Thu, 27 Oct 2011)*
228 This was the fourth release candidate of the 2.5 series.
234 *(Released Wed, 26 Oct 2011)*
236 This was the third release candidate of the 2.5 series.
242 *(Released Tue, 18 Oct 2011)*
244 This was the second release candidate of the 2.5 series.
250 *(Released Tue, 4 Oct 2011)*
252 This was the first release candidate of the 2.5 series.
258 *(Released Wed, 31 Aug 2011)*
260 This was the third beta release of the 2.5 series.
266 *(Released Mon, 22 Aug 2011)*
268 This was the second beta release of the 2.5 series.
274 *(Released Fri, 12 Aug 2011)*
276 This was the first beta release of the 2.5 series.
282 *(Released Thu, 27 Oct 2011)*
284 - Fixed bug when parsing command line parameter values ending in
286 - Fixed assertion error after unclean master shutdown
287 - Disable HTTP client pool for RPC, significantly reducing memory usage
289 - Fixed queue archive creation with wrong permissions
295 *(Released Tue, 23 Aug 2011)*
299 - Fixed documentation for importing with ``--src-dir`` option
300 - Fixed a bug in ``ensure-dirs`` with queue/archive permissions
301 - Fixed a parsing issue with DRBD 8.3.11 in the Linux kernel
307 *(Released Fri, 5 Aug 2011)*
309 Many bug-fixes and a few small features:
311 - Fixed argument order in ``ReserveLV`` and ``ReserveMAC`` which caused
312 issues when you tried to add an instance with two MAC addresses in one
314 - KVM: fixed per-instance stored UID value
315 - KVM: configure bridged NICs at migration start
316 - KVM: Fix a bug where instance will not start with never KVM versions
318 - Added OS search path to ``gnt-cluster info``
319 - Fixed an issue with ``file_storage_dir`` where you were forced to
320 provide an absolute path, but the documentation states it is a
321 relative path, the documentation was right
322 - Added a new parameter to instance stop/start called ``--no-remember``
323 that will make the state change to not be remembered
324 - Implemented ``no_remember`` at RAPI level
325 - Improved the documentation
326 - Node evacuation: don't call IAllocator if node is already empty
327 - Fixed bug in DRBD8 replace disks on current nodes
328 - Fixed bug in recreate-disks for DRBD instances
329 - Moved assertion checking locks in ``gnt-instance replace-disks``
330 causing it to abort with not owning the right locks for some situation
331 - Job queue: Fixed potential race condition when cancelling queued jobs
332 - Fixed off-by-one bug in job serial generation
333 - ``gnt-node volumes``: Fix instance names
334 - Fixed aliases in bash completion
335 - Fixed a bug in reopening log files after being sent a SIGHUP
336 - Added a flag to burnin to allow specifying VCPU count
337 - Bugfixes to non-root Ganeti configuration
343 *(Released Thu, 12 May 2011)*
345 Many bug-fixes and a few new small features:
347 - Fixed a bug related to log opening failures
348 - Fixed a bug in instance listing with orphan instances
349 - Fixed a bug which prevented resetting the cluster-level node parameter
350 ``oob_program`` to the default
351 - Many fixes related to the ``cluster-merge`` tool
352 - Fixed a race condition in the lock monitor, which caused failures
353 during (at least) creation of many instances in parallel
354 - Improved output for gnt-job info
355 - Removed the quiet flag on some ssh calls which prevented debugging
357 - Improved the N+1 failure messages in cluster verify by actually
358 showing the memory values (needed and available)
359 - Increased lock attempt timeouts so that when executing long operations
360 (e.g. DRBD replace-disks) other jobs do not enter 'blocking acquire'
361 too early and thus prevent the use of the 'fair' mechanism
362 - Changed instance query data (``gnt-instance info``) to not acquire
363 locks unless needed, thus allowing its use on locked instance if only
364 static information is asked for
365 - Improved behaviour with filesystems that do not support rename on an
367 - Fixed the behaviour of ``prealloc_wipe_disks`` cluster parameter which
368 kept locks on all nodes during the wipe, which is unneeded
369 - Fixed ``gnt-watcher`` handling of errors during hooks execution
370 - Fixed bug in ``prealloc_wipe_disks`` with small disk sizes (less than
371 10GiB) which caused the wipe to fail right at the end in some cases
372 - Fixed master IP activation when doing master failover with no-voting
373 - Fixed bug in ``gnt-node add --readd`` which allowed the re-adding of
374 the master node itself
375 - Fixed potential data-loss in under disk full conditions, where Ganeti
376 wouldn't check correctly the return code and would consider
377 partially-written files 'correct'
378 - Fixed bug related to multiple VGs and DRBD disk replacing
379 - Added new disk parameter ``metavg`` that allows placement of the meta
380 device for DRBD in a different volume group
381 - Fixed error handling in the node daemon when the system libc doesn't
382 have major number 6 (i.e. if ``libc.so.6`` is not the actual libc)
383 - Fixed lock release during replace-disks, which kept cluster-wide locks
384 when doing disk replaces with an iallocator script
385 - Added check for missing bridges in cluster verify
386 - Handle EPIPE errors while writing to the terminal better, so that
387 piping the output to e.g. ``less`` doesn't cause a backtrace
388 - Fixed rare case where a ^C during Luxi calls could have been
389 interpreted as server errors, instead of simply terminating
390 - Fixed a race condition in LUGroupAssignNodes (``gnt-group
392 - Added a few more parameters to the KVM hypervisor, allowing a second
393 CDROM, custom disk type for CDROMs and a floppy image
394 - Removed redundant message in instance rename when the name is given
396 - Added option to ``gnt-instance recreate-disks`` to allow creating the
397 disks on new nodes, allowing recreation when the original instance
398 nodes are completely gone
399 - Added option when converting disk templates to DRBD to skip waiting
400 for the resync, in order to make the instance available sooner
401 - Added two new variables to the OS scripts environment (containing the
403 - Made the root_path and optional parameter for the xen-pvm hypervisor,
404 to allow use of ``pvgrub`` as bootloader
405 - Changed the instance memory modifications to only check out-of-memory
406 conditions on memory increases, and turned the secondary node warnings
407 into errors (they can still be overridden via ``--force``)
408 - Fixed the handling of a corner case when the Python installation gets
409 corrupted (e.g. a bad disk) while ganeti-noded is running and we try
410 to execute a command that doesn't exist
411 - Fixed a bug in ``gnt-instance move`` (LUInstanceMove) when the primary
412 node of the instance returned failures during instance shutdown; this
413 adds the option ``--ignore-consistency`` to gnt-instance move
415 And as usual, various improvements to the error messages, documentation
422 *(Released Wed, 09 Mar 2011)*
424 Emergency bug-fix release. ``tools/cfgupgrade`` was broken and overwrote
425 the RAPI users file if run twice (even with ``--dry-run``).
427 The release fixes that bug (nothing else changed).
433 *(Released Mon, 07 Mar 2011)*
435 Final 2.4.0 release. Just a few small fixes:
437 - Fixed RAPI node evacuate
438 - Fixed the kvm-ifup script
439 - Fixed internal error handling for special job cases
440 - Updated man page to specify the escaping feature for options
446 *(Released Mon, 28 Feb 2011)*
448 A critical fix for the ``prealloc_wipe_disks`` feature: it is possible
449 that this feature wiped the disks of the wrong instance, leading to loss
454 - Fixed title of query field containing instance name
455 - Expanded the glossary in the documentation
456 - Fixed one unittest (internal issue)
462 *(Released Mon, 21 Feb 2011)*
464 A number of bug fixes plus just a couple functionality changes.
466 On the user-visible side, the ``gnt-* list`` command output has changed
467 with respect to "special" field states. The current rc1 style of display
468 can be re-enabled by passing a new ``--verbose`` (``-v``) flag, but in
469 the default output mode special fields states are displayed as follows:
471 - Offline resource: ``*``
472 - Unavailable/not applicable: ``-``
473 - Data missing (RPC failure): ``?``
474 - Unknown field: ``??``
476 Another user-visible change is the addition of ``--force-join`` to
481 - ``tools/cluster-merge`` has seen many fixes and is now enabled again
482 - Fixed regression in RAPI/instance reinstall where all parameters were
483 required (instead of optional)
484 - Fixed ``gnt-cluster repair-disk-sizes``, was broken since Ganeti 2.2
485 - Fixed iallocator usage (offline nodes were not considered offline)
486 - Fixed ``gnt-node list`` with respect to non-vm_capable nodes
487 - Fixed hypervisor and OS parameter validation with respect to
489 - Fixed ``gnt-cluster verify`` with respect to offline nodes (mostly
491 - Fixed ``tools/listrunner`` with respect to agent-based usage
497 *(Released Fri, 4 Feb 2011)*
499 Many changes and fixes since the beta1 release. While there were some
500 internal changes, the code has been mostly stabilised for the RC
503 Note: the dumb allocator was removed in this release, as it was not kept
504 up-to-date with the IAllocator protocol changes. It is recommended to
505 use the ``hail`` command from the ganeti-htools package.
507 Note: the 2.4 and up versions of Ganeti are not compatible with the
508 0.2.x branch of ganeti-htools. You need to upgrade to
509 ganeti-htools-0.3.0 (or later).
511 Regressions fixed from 2.3
512 ~~~~~~~~~~~~~~~~~~~~~~~~~~
514 - Fixed the ``gnt-cluster verify-disks`` command
515 - Made ``gnt-cluster verify-disks`` work in parallel (as opposed to
517 - Fixed disk adoption breakage
518 - Fixed wrong headers in instance listing for field aliases
523 - Fixed corner case in KVM handling of NICs
524 - Fixed many cases of wrong handling of non-vm_capable nodes
525 - Fixed a bug where a missing instance symlink was not possible to
526 recreate with any ``gnt-*`` command (now ``gnt-instance
527 activate-disks`` does it)
528 - Fixed the volume group name as reported by ``gnt-cluster
530 - Increased timeouts for the import-export code, hopefully leading to
531 fewer aborts due network or instance timeouts
532 - Fixed bug in ``gnt-node list-storage``
533 - Fixed bug where not all daemons were started on cluster
534 initialisation, but only at the first watcher run
535 - Fixed many bugs in the OOB implementation
536 - Fixed watcher behaviour in presence of instances with offline
538 - Fixed instance list output for instances running on the wrong node
539 - a few fixes to the cluster-merge tool, but it still cannot merge
540 multi-node groups (currently it is not recommended to use this tool)
546 - Improved network configuration for the KVM hypervisor
547 - Added e1000 as a supported NIC for Xen-HVM
548 - Improved the lvmstrap tool to also be able to use partitions, as
549 opposed to full disks
550 - Improved speed of disk wiping (the cluster parameter
551 ``prealloc_wipe_disks``, so that it has a low impact on the total time
552 of instance creations
553 - Added documentation for the OS parameters
554 - Changed ``gnt-instance deactivate-disks`` so that it can work if the
555 hypervisor is not responding
556 - Added display of blacklisted and hidden OS information in
558 - Extended ``gnt-cluster verify`` to also validate hypervisor, backend,
559 NIC and node parameters, which might create problems with currently
560 invalid (but undetected) configuration files, but prevents validation
561 failures when unrelated parameters are modified
562 - Changed cluster initialisation to wait for the master daemon to become
564 - Expanded the RAPI interface:
566 - Added config redistribution resource
567 - Added activation/deactivation of instance disks
568 - Added export of console information
570 - Implemented log file reopening on SIGHUP, which allows using
571 logrotate(8) for the Ganeti log files
572 - Added a basic OOB helper script as an example
578 *(Released Fri, 14 Jan 2011)*
583 - Fixed timezone issues when formatting timestamps
584 - Added support for node groups, available via ``gnt-group`` and other
586 - Added out-of-band framework and management, see :doc:`design
587 document <design-oob>`
588 - Removed support for roman numbers from ``gnt-node list`` and
589 ``gnt-instance list``.
590 - Allowed modification of master network interface via ``gnt-cluster
591 modify --master-netdev``
592 - Accept offline secondaries while shutting down instance disks
593 - Added ``blockdev_prefix`` parameter to Xen PVM and HVM hypervisors
594 - Added support for multiple LVM volume groups
595 - Avoid sorting nodes for ``gnt-node list`` if specific nodes are
597 - Added commands to list available fields:
599 - ``gnt-node list-fields``
600 - ``gnt-group list-fields``
601 - ``gnt-instance list-fields``
603 - Updated documentation and man pages
608 - Moved ``rapi_users`` file into separate directory, now named
609 ``.../ganeti/rapi/users``, ``cfgupgrade`` moves the file and creates a
611 - Added new tool for running commands on many machines,
612 ``tools/ganeti-listrunner``
613 - Implemented more verbose result in ``OpInstanceConsole`` opcode, also
614 improving the ``gnt-instance console`` output
615 - Allowed customisation of disk index separator at ``configure`` time
616 - Export node group allocation policy to :doc:`iallocator <iallocator>`
617 - Added support for non-partitioned md disks in ``lvmstrap``
618 - Added script to gracefully power off KVM instances
619 - Split ``utils`` module into smaller parts
620 - Changed query operations to return more detailed information, e.g.
621 whether an information is unavailable due to an offline node. To use
622 this new functionality, the LUXI call ``Query`` must be used. Field
623 information is now stored by the master daemon and can be retrieved
624 using ``QueryFields``. Instances, nodes and groups can also be queried
625 using the new opcodes ``OpQuery`` and ``OpQueryFields`` (not yet
626 exposed via RAPI). The following commands make use of this
627 infrastructure change:
630 - ``gnt-group list-fields``
632 - ``gnt-node list-fields``
633 - ``gnt-instance list``
634 - ``gnt-instance list-fields``
635 - ``gnt-debug locks``
640 - New RAPI resources (see :doc:`rapi`):
644 - ``/2/groups/[group_name]``
645 - ``/2/groups/[group_name]/assign-nodes``
646 - ``/2/groups/[group_name]/modify``
647 - ``/2/groups/[group_name]/rename``
648 - ``/2/instances/[instance_name]/disk/[disk_index]/grow``
652 - Implemented ``no_install`` for instance creation
653 - Implemented OS parameters for instance reinstallation, allowing
654 use of special settings on reinstallation (e.g. for preserving data)
659 - Added IPv6 support in import/export
660 - Pause DRBD synchronization while wiping disks on instance creation
661 - Updated unittests and QA scripts
662 - Improved network parameters passed to KVM
663 - Converted man pages from docbook to reStructuredText
669 *(Released Mon, 20 Dec 2010)*
671 Released version 2.3.1~rc1 without any changes.
677 *(Released Wed, 1 Dec 2010)*
679 - impexpd: Disable OpenSSL compression in socat if possible (backport
680 from master, commit e90739d625b, see :doc:`installation guide
681 <install-quick>` for details)
682 - Changed unittest coverage report to exclude test scripts
683 - Added script to check version format
689 *(Released Wed, 1 Dec 2010)*
691 Released version 2.3.0~rc1 without any changes.
697 *(Released Fri, 19 Nov 2010)*
699 A number of bugfixes and documentation updates:
701 - Update ganeti-os-interface documentation
702 - Fixed a bug related to duplicate MACs or similar items which should be
704 - Fix breakage in OS state modify
705 - Reinstall instance: disallow offline secondaries (fixes bug related to
706 OS changing but reinstall failing)
707 - plus all the other fixes between 2.2.1 and 2.2.2
713 *(Released Tue, 2 Nov 2010)*
715 - Fixed clearing of the default iallocator using ``gnt-cluster modify``
716 - Fixed master failover race with watcher
717 - Fixed a bug in ``gnt-node modify`` which could lead to an inconsistent
719 - Accept previously stopped instance for export with instance removal
720 - Simplify and extend the environment variables for instance OS scripts
721 - Added new node flags, ``master_capable`` and ``vm_capable``
722 - Added optional instance disk wiping prior during allocation. This is a
723 cluster-wide option and can be set/modified using
724 ``gnt-cluster {init,modify} --prealloc-wipe-disks``.
725 - Added IPv6 support, see :doc:`design document <design-2.3>` and
727 - Added a new watcher option (``--ignore-pause``)
728 - Added option to ignore offline node on instance start/stop
729 (``--ignore-offline``)
730 - Allow overriding OS parameters with ``gnt-instance reinstall``
731 - Added ability to change node's secondary IP address using ``gnt-node
733 - Implemented privilege separation for all daemons except
734 ``ganeti-noded``, see ``configure`` options
735 - Complain if an instance's disk is marked faulty in ``gnt-cluster
737 - Implemented job priorities (see ``ganeti(7)`` manpage)
738 - Ignore failures while shutting down instances during failover from
740 - Exit daemon's bootstrap process only once daemon is ready
741 - Export more information via ``LUInstanceQuery``/remote API
742 - Improved documentation, QA and unittests
743 - RAPI daemon now watches ``rapi_users`` all the time and doesn't need a
744 restart if the file was created or changed
745 - Added LUXI protocol version sent with each request and response,
746 allowing detection of server/client mismatches
747 - Moved the Python scripts among gnt-* and ganeti-* into modules
748 - Moved all code related to setting up SSH to an external script,
750 - Infrastructure changes for node group support in future versions
756 *(Released Fri, 19 Nov 2010)*
758 A few small bugs fixed, and some improvements to the build system:
760 - Fix documentation regarding conversion to drbd
761 - Fix validation of parameters in cluster modify (``gnt-cluster modify
763 - Fix error handling in node modify with multiple changes
764 - Allow remote imports without checked names
770 *(Released Tue, 19 Oct 2010)*
772 - Disable SSL session ID cache in RPC client
778 *(Released Thu, 14 Oct 2010)*
780 - Fix interaction between Curl/GnuTLS and the Python's HTTP server
781 (thanks Apollon Oikonomopoulos!), finally allowing the use of Curl
783 - Fix problems with interaction between Curl and Python's HTTP server,
784 resulting in increased speed in many RPC calls
785 - Improve our release script to prevent breakage with older aclocal and
792 *(Released Thu, 7 Oct 2010)*
794 - Fixed issue 125, replace hardcoded "xenvg" in ``gnt-cluster`` with
795 value retrieved from master
796 - Added support for blacklisted or hidden OS definitions
797 - Added simple lock monitor (accessible via (``gnt-debug locks``)
798 - Added support for -mem-path in KVM hypervisor abstraction layer
799 - Allow overriding instance parameters in tool for inter-cluster
800 instance moves (``tools/move-instance``)
801 - Improved opcode summaries (e.g. in ``gnt-job list``)
802 - Improve consistency of OS listing by sorting it
803 - Documentation updates
809 *(Released Fri, 8 Oct 2010)*
811 - Rebuild with a newer autotools version, to fix python 2.6 compatibility
817 *(Released Mon, 4 Oct 2010)*
819 - Fixed regression in ``gnt-instance rename``
825 *(Released Wed, 22 Sep 2010)*
827 - Fixed OS_VARIANT variable for OS scripts
828 - Fixed cluster tag operations via RAPI
829 - Made ``setup-ssh`` exit with non-zero code if an error occurred
830 - Disabled RAPI CA checks in watcher
836 *(Released Mon, 23 Aug 2010)*
838 - Support DRBD versions of the format "a.b.c.d"
840 - Re-introduce support for usage from multiple threads in RAPI client
841 - Instance renames and modify via RAPI
842 - Work around race condition between processing and archival in job
844 - Mark opcodes following failed one as failed, too
845 - Job field ``lock_status`` was removed due to difficulties making it
846 work with the changed job queue in Ganeti 2.2; a better way to monitor
847 locks is expected for a later 2.2.x release
848 - Fixed dry-run behaviour with many commands
849 - Support ``ssh-agent`` again when adding nodes
850 - Many additional bugfixes
856 *(Released Fri, 30 Jul 2010)*
858 Important change: the internal RPC mechanism between Ganeti nodes has
859 changed from using a home-grown http library (based on the Python base
860 libraries) to use the PycURL library. This requires that PycURL is
861 installed on nodes. Please note that on Debian/Ubuntu, PycURL is linked
862 against GnuTLS by default. cURL's support for GnuTLS had known issues
863 before cURL 7.21.0 and we recommend using the latest cURL release or
864 linking against OpenSSL. Most other distributions already link PycURL
865 and cURL against OpenSSL. The command::
867 python -c 'import pycurl; print pycurl.version'
869 can be used to determine the libraries PycURL and cURL are linked
872 Other significant changes:
874 - Rewrote much of the internals of the job queue, in order to achieve
875 better parallelism; this decouples job query operations from the job
876 processing, and it should allow much nicer behaviour of the master
877 daemon under load, and it also has uncovered some long-standing bugs
878 related to the job serialisation (now fixed)
879 - Added a default iallocator setting to the cluster parameters,
880 eliminating the need to always pass nodes or an iallocator for
881 operations that require selection of new node(s)
882 - Added experimental support for the LXC virtualization method
883 - Added support for OS parameters, which allows the installation of
884 instances to pass parameter to OS scripts in order to customise the
886 - Added a hypervisor parameter controlling the migration type (live or
887 non-live), since hypervisors have various levels of reliability; this
888 has renamed the 'live' parameter to 'mode'
889 - Added a cluster parameter ``reserved_lvs`` that denotes reserved
890 logical volumes, meaning that cluster verify will ignore them and not
891 flag their presence as errors
892 - The watcher will now reset the error count for failed instances after
893 8 hours, thus allowing self-healing if the problem that caused the
894 instances to be down/fail to start has cleared in the meantime
895 - Added a cluster parameter ``drbd_usermode_helper`` that makes Ganeti
896 check for, and warn, if the drbd module parameter ``usermode_helper``
897 is not consistent with the cluster-wide setting; this is needed to
898 make diagnose easier of failed drbd creations
899 - Started adding base IPv6 support, but this is not yet
900 enabled/available for use
901 - Rename operations (cluster, instance) will now return the new name,
902 which is especially useful if a short name was passed in
903 - Added support for instance migration in RAPI
904 - Added a tool to pre-configure nodes for the SSH setup, before joining
905 them to the cluster; this will allow in the future a simplified model
906 for node joining (but not yet fully enabled in 2.2); this needs the
907 paramiko python library
908 - Fixed handling of name-resolving errors
909 - Fixed consistency of job results on the error path
910 - Fixed master-failover race condition when executed multiple times in
912 - Fixed many bugs related to the job queue (mostly introduced during the
913 2.2 development cycle, so not all are impacting 2.1)
914 - Fixed instance migration with missing disk symlinks
915 - Fixed handling of unknown jobs in ``gnt-job archive``
916 - And many other small fixes/improvements
920 - Enhanced both the unittest and the QA coverage
921 - Switched the opcode validation to a generic model, and extended the
922 validation to all opcode parameters
923 - Changed more parts of the code that write shell scripts to use the
925 - Switched the master daemon to use the asyncore library for the Luxi
932 *(Released Thu, 17 Jun 2010)*
934 - Added tool (``move-instance``) and infrastructure to move instances
935 between separate clusters (see :doc:`separate documentation
936 <move-instance>` and :doc:`design document <design-2.2>`)
937 - Added per-request RPC timeout
938 - RAPI now requires a Content-Type header for requests with a body (e.g.
939 ``PUT`` or ``POST``) which must be set to ``application/json`` (see
940 :rfc:`2616` (HTTP/1.1), section 7.2.1)
941 - ``ganeti-watcher`` attempts to restart ``ganeti-rapi`` if RAPI is not
943 - Implemented initial support for running Ganeti daemons as separate
944 users, see configure-time flags ``--with-user-prefix`` and
945 ``--with-group-prefix`` (only ``ganeti-rapi`` is supported at this
947 - Instances can be removed after export (``gnt-backup export
949 - Self-signed certificates generated by Ganeti now use a 2048 bit RSA
950 key (instead of 1024 bit)
951 - Added new cluster configuration file for cluster domain secret
952 - Import/export now use SSL instead of SSH
953 - Added support for showing estimated time when exporting an instance,
954 see the ``ganeti-os-interface(7)`` manpage and look for
961 *(Released Tue, 16 Nov 2010)*
963 Some more bugfixes. Unless critical bugs occur, this will be the last
966 - Fix case of MAC special-values
967 - Fix mac checker regex
968 - backend: Fix typo causing "out of range" error
969 - Add missing --units in gnt-instance list man page
975 *(Released Tue, 24 Aug 2010)*
978 - Don't ignore secondary node silently on non-mirrored disk templates
980 - Fix --master-netdev arg name in gnt-cluster(8) (issue 114)
981 - Fix usb_mouse parameter breaking with vnc_console (issue 109)
982 - Properly document the usb_mouse parameter
983 - Fix path in ganeti-rapi(8) (issue 116)
984 - Adjust error message when the ganeti user's .ssh directory is
986 - Add same-node-check when changing the disk template to drbd
992 *(Released Fri, 16 Jul 2010)*
995 - Add an option to only select some reboot types during qa/burnin.
996 (on some hypervisors consequent reboots are not supported)
997 - Fix infrequent race condition in master failover. Sometimes the old
998 master ip address would be still detected as up for a short time
999 after it was removed, causing failover to fail.
1000 - Decrease mlockall warnings when the ctypes module is missing. On
1001 Python 2.4 we support running even if no ctypes module is installed,
1002 but we were too verbose about this issue.
1003 - Fix building on old distributions, on which man doesn't have a
1005 - Fix RAPI not to ignore the MAC address on instance creation
1006 - Implement the old instance creation format in the RAPI client.
1012 *(Released Thu, 01 Jul 2010)*
1014 A small bugfix release:
1015 - Fix disk adoption: broken by strict --disk option checking in 2.1.4
1016 - Fix batch-create: broken in the whole 2.1 series due to a lookup on
1017 a non-existing option
1018 - Fix instance create: the --force-variant option was ignored
1019 - Improve pylint 0.21 compatibility and warnings with Python 2.6
1020 - Fix modify node storage with non-FQDN arguments
1021 - Fix RAPI client to authenticate under Python 2.6 when used
1022 for more than 5 requests needing authentication
1023 - Fix gnt-instance modify -t (storage) giving a wrong error message
1024 when converting a non-shutdown drbd instance to plain
1030 *(Released Fri, 18 Jun 2010)*
1032 A small bugfix release:
1034 - Fix live migration of KVM instances started with older Ganeti
1035 versions which had fewer hypervisor parameters
1036 - Fix gnt-instance grow-disk on down instances
1037 - Fix an error-reporting bug during instance migration
1038 - Better checking of the ``--net`` and ``--disk`` values, to avoid
1039 silently ignoring broken ones
1040 - Fix an RPC error reporting bug affecting, for example, RAPI client
1042 - Fix bug triggered by different API version os-es on different nodes
1043 - Fix a bug in instance startup with custom hvparams: OS level
1044 parameters would fail to be applied.
1045 - Fix the RAPI client under Python 2.6 (but more work is needed to
1046 make it work completely well with OpenSSL)
1047 - Fix handling of errors when resolving names from DNS
1053 *(Released Thu, 3 Jun 2010)*
1055 A medium sized development cycle. Some new features, and some
1056 fixes/small improvements/cleanups.
1058 Significant features
1059 ~~~~~~~~~~~~~~~~~~~~
1061 The node deamon now tries to mlock itself into memory, unless the
1062 ``--no-mlock`` flag is passed. It also doesn't fail if it can't write
1063 its logs, and falls back to console logging. This allows emergency
1064 features such as ``gnt-node powercycle`` to work even in the event of a
1065 broken node disk (tested offlining the disk hosting the node's
1066 filesystem and dropping its memory caches; don't try this at home)
1068 KVM: add vhost-net acceleration support. It can be tested with a new
1069 enough version of the kernel and of qemu-kvm.
1071 KVM: Add instance chrooting feature. If you use privilege dropping for
1072 your VMs you can also now force them to chroot to an empty directory,
1073 before starting the emulated guest.
1075 KVM: Add maximum migration bandwith and maximum downtime tweaking
1076 support (requires a new-enough version of qemu-kvm).
1078 Cluster verify will now warn if the master node doesn't have the master
1079 ip configured on it.
1081 Add a new (incompatible) instance creation request format to RAPI which
1082 supports all parameters (previously only a subset was supported, and it
1083 wasn't possible to extend the old format to accomodate all the new
1084 features. The old format is still supported, and a client can check for
1085 this feature, before using it, by checking for its presence in the
1086 ``features`` RAPI resource.
1088 Now with ancient latin support. Try it passing the ``--roman`` option to
1089 ``gnt-instance info``, ``gnt-cluster info`` or ``gnt-node list``
1090 (requires the python-roman module to be installed, in order to work).
1095 As usual many internal code refactorings, documentation updates, and
1098 - Lots of improvements and cleanups to the experimental Remote API
1099 (RAPI) client library.
1100 - A new unit test suite for the core daemon libraries.
1101 - A fix to creating missing directories makes sure the umask is not
1102 applied anymore. This enforces the same directory permissions
1104 - Better handling terminating daemons with ctrl+c (used when running
1105 them in debugging mode).
1106 - Fix a race condition in live migrating a KVM instance, when stat()
1107 on the old proc status file returned EINVAL, which is an unexpected
1109 - Fixed manpage checking with newer man and utf-8 charachters. But now
1110 you need the en_US.UTF-8 locale enabled to build Ganeti from git.
1116 *(Released Fri, 7 May 2010)*
1118 Fix a bug which prevented untagged KVM instances from starting.
1124 *(Released Fri, 7 May 2010)*
1126 Another release with a long development cycle, during which many
1127 different features were added.
1129 Significant features
1130 ~~~~~~~~~~~~~~~~~~~~
1132 The KVM hypervisor now can run the individual instances as non-root, to
1133 reduce the impact of a VM being hijacked due to bugs in the
1134 hypervisor. It is possible to run all instances as a single (non-root)
1135 user, to manually specify a user for each instance, or to dynamically
1136 allocate a user out of a cluster-wide pool to each instance, with the
1137 guarantee that no two instances will run under the same user ID on any
1140 An experimental RAPI client library, that can be used standalone
1141 (without the other Ganeti libraries), is provided in the source tree as
1142 ``lib/rapi/client.py``. Note this client might change its interface in
1143 the future, as we iterate on its capabilities.
1145 A new command, ``gnt-cluster renew-crypto`` has been added to easily
1146 replace the cluster's certificates and crypto keys. This might help in
1147 case they have been compromised, or have simply expired.
1149 A new disk option for instance creation has been added that allows one
1150 to "adopt" currently existing logical volumes, with data
1151 preservation. This should allow easier migration to Ganeti from
1152 unmanaged (or managed via other software) instances.
1154 Another disk improvement is the possibility to convert between redundant
1155 (DRBD) and plain (LVM) disk configuration for an instance. This should
1156 allow better scalability (starting with one node and growing the
1157 cluster, or shrinking a two-node cluster to one node).
1159 A new feature that could help with automated node failovers has been
1160 implemented: if a node sees itself as offline (by querying the master
1161 candidates), it will try to shutdown (hard) all instances and any active
1162 DRBD devices. This reduces the risk of duplicate instances if an
1163 external script automatically failovers the instances on such nodes. To
1164 enable this, the cluster parameter ``maintain_node_health`` should be
1165 enabled; in the future this option (per the name) will enable other
1166 automatic maintenance features.
1168 Instance export/import now will reuse the original instance
1169 specifications for all parameters; that means exporting an instance,
1170 deleting it and the importing it back should give an almost identical
1171 instance. Note that the default import behaviour has changed from
1172 before, where it created only one NIC; now it recreates the original
1175 Cluster verify has added a few new checks: SSL certificates validity,
1176 /etc/hosts consistency across the cluster, etc.
1181 As usual, many internal changes were done, documentation fixes,
1184 - Fixed cluster initialization with disabled cluster storage (regression
1185 introduced in 2.1.1)
1186 - File-based storage supports growing the disks
1187 - Fixed behaviour of node role changes
1188 - Fixed cluster verify for some corner cases, plus a general rewrite of
1189 cluster verify to allow future extension with more checks
1190 - Fixed log spamming by watcher and node daemon (regression introduced
1192 - Fixed possible validation issues when changing the list of enabled
1194 - Fixed cleanup of /etc/hosts during node removal
1195 - Fixed RAPI response for invalid methods
1196 - Fixed bug with hashed passwords in ``ganeti-rapi`` daemon
1197 - Multiple small improvements to the KVM hypervisor (VNC usage, booting
1198 from ide disks, etc.)
1199 - Allow OS changes without re-installation (to record a changed OS
1200 outside of Ganeti, or to allow OS renames)
1201 - Allow instance creation without OS installation (useful for example if
1202 the OS will be installed manually, or restored from a backup not in
1204 - Implemented option to make cluster ``copyfile`` use the replication
1206 - Added list of enabled hypervisors to ssconf (possibly useful for
1208 - Added a new tool (``tools/cfgupgrade12``) that allows upgrading from
1210 - A partial form of node re-IP is possible via node readd, which now
1211 allows changed node primary IP
1212 - Command line utilities now show an informational message if the job is
1214 - The logs of the master daemon now show the PID/UID/GID of the
1221 *(Released Fri, 12 Mar 2010)*
1223 During the 2.1.0 long release candidate cycle, a lot of improvements and
1224 changes have accumulated with were released later as 2.1.1.
1229 The node evacuate command (``gnt-node evacuate``) was significantly
1230 rewritten, and as such the IAllocator protocol was changed - a new
1231 request type has been added. This unfortunate change during a stable
1232 series is designed to improve performance of node evacuations; on
1233 clusters with more than about five nodes and which are well-balanced,
1234 evacuation should proceed in parallel for all instances of the node
1235 being evacuated. As such, any existing IAllocator scripts need to be
1236 updated, otherwise the above command will fail due to the unknown
1237 request. The provided "dumb" allocator has not been updated; but the
1238 ganeti-htools package supports the new protocol since version 0.2.4.
1240 Another important change is increased validation of node and instance
1241 names. This might create problems in special cases, if invalid host
1242 names are being used.
1244 Also, a new layer of hypervisor parameters has been added, that sits at
1245 OS level between the cluster defaults and the instance ones. This allows
1246 customisation of virtualization parameters depending on the installed
1247 OS. For example instances with OS 'X' may have a different KVM kernel
1248 (or any other parameter) than the cluster defaults. This is intended to
1249 help managing a multiple OSes on the same cluster, without manual
1250 modification of each instance's parameters.
1252 A tool for merging clusters, ``cluster-merge``, has been added in the
1253 tools sub-directory.
1258 - Improved the int/float conversions that should make the code more
1259 robust in face of errors from the node daemons
1260 - Fixed the remove node code in case of internal configuration errors
1261 - Fixed the node daemon behaviour in face of inconsistent queue
1262 directory (e.g. read-only file-system where we can't open the files
1264 - Fixed the behaviour of gnt-node modify for master candidate demotion;
1265 now it either aborts cleanly or, if given the new "auto_promote"
1266 parameter, will automatically promote other nodes as needed
1267 - Fixed compatibility with (unreleased yet) Python 2.6.5 that would
1268 completely prevent Ganeti from working
1269 - Fixed bug for instance export when not all disks were successfully
1271 - Fixed behaviour of node add when the new node is slow in starting up
1273 - Fixed handling of signals in the LUXI client, which should improve
1274 behaviour of command-line scripts
1275 - Added checks for invalid node/instance names in the configuration (now
1276 flagged during cluster verify)
1277 - Fixed watcher behaviour for disk activation errors
1278 - Fixed two potentially endless loops in http library, which led to the
1279 RAPI daemon hanging and consuming 100% CPU in some cases
1280 - Fixed bug in RAPI daemon related to hashed passwords
1281 - Fixed bug for unintended qemu-level bridging of multi-NIC KVM
1283 - Enhanced compatibility with non-Debian OSes, but not using absolute
1284 path in some commands and allowing customisation of the ssh
1285 configuration directory
1286 - Fixed possible future issue with new Python versions by abiding to the
1287 proper use of ``__slots__`` attribute on classes
1288 - Added checks that should prevent directory traversal attacks
1289 - Many documentation fixes based on feedback from users
1294 - Added an "early_release" more for instance replace disks and node
1295 evacuate, where we release locks earlier and thus allow higher
1296 parallelism within the cluster
1297 - Added watcher hooks, intended to allow the watcher to restart other
1298 daemons (e.g. from the ganeti-nbma project), but they can be used of
1299 course for any other purpose
1300 - Added a compile-time disable for DRBD barriers, to increase
1301 performance if the administrator trusts the power supply or the
1302 storage system to not lose writes
1303 - Added the option of using syslog for logging instead of, or in
1304 addition to, Ganeti's own log files
1305 - Removed boot restriction for paravirtual NICs for KVM, recent versions
1306 can indeed boot from a paravirtual NIC
1307 - Added a generic debug level for many operations; while this is not
1308 used widely yet, it allows one to pass the debug value all the way to
1310 - Enhanced the hooks environment for instance moves (failovers,
1311 migrations) where the primary/secondary nodes changed during the
1312 operation, by adding {NEW,OLD}_{PRIMARY,SECONDARY} vars
1313 - Enhanced data validations for many user-supplied values; one important
1314 item is the restrictions imposed on instance and node names, which
1315 might reject some (invalid) host names
1316 - Add a configure-time option to disable file-based storage, if it's not
1317 needed; this allows greater security separation between the master
1318 node and the other nodes from the point of view of the inter-node RPC
1320 - Added user notification in interactive tools if job is waiting in the
1321 job queue or trying to acquire locks
1322 - Added log messages when a job is waiting for locks
1323 - Added filtering by node tags in instance operations which admit
1324 multiple instances (start, stop, reboot, reinstall)
1325 - Added a new tool for cluster mergers, ``cluster-merge``
1326 - Parameters from command line which are of the form ``a=b,c=d`` can now
1327 use backslash escapes to pass in values which contain commas,
1328 e.g. ``a=b\\c,d=e`` where the 'a' parameter would get the value
1330 - For KVM, the instance name is the first parameter passed to KVM, so
1331 that it's more visible in the process list
1337 *(Released Tue, 2 Mar 2010)*
1339 Ganeti 2.1 brings many improvements with it. Major changes:
1341 - Added infrastructure to ease automated disk repairs
1342 - Added new daemon to export configuration data in a cheaper way than
1343 using the remote API
1344 - Instance NICs can now be routed instead of being associated with a
1346 - Improved job locking logic to reduce impact of jobs acquiring multiple
1347 locks waiting for other long-running jobs
1349 In-depth implementation details can be found in the Ganeti 2.1 design
1355 - Added chroot hypervisor
1356 - Added more options to xen-hvm hypervisor (``kernel_path`` and
1358 - Added more options to xen-pvm hypervisor (``use_bootloader``,
1359 ``bootloader_path`` and ``bootloader_args``)
1360 - Added the ``use_localtime`` option for the xen-hvm and kvm
1361 hypervisors, and the default value for this has changed to false (in
1362 2.0 xen-hvm always enabled it)
1363 - Added luxi call to submit multiple jobs in one go
1364 - Added cluster initialization option to not modify ``/etc/hosts``
1366 - Added network interface parameters
1367 - Added dry run mode to some LUs
1368 - Added RAPI resources:
1370 - ``/2/instances/[instance_name]/info``
1371 - ``/2/instances/[instance_name]/replace-disks``
1372 - ``/2/nodes/[node_name]/evacuate``
1373 - ``/2/nodes/[node_name]/migrate``
1374 - ``/2/nodes/[node_name]/role``
1375 - ``/2/nodes/[node_name]/storage``
1376 - ``/2/nodes/[node_name]/storage/modify``
1377 - ``/2/nodes/[node_name]/storage/repair``
1379 - Added OpCodes to evacuate or migrate all instances on a node
1380 - Added new command to list storage elements on nodes (``gnt-node
1381 list-storage``) and modify them (``gnt-node modify-storage``)
1382 - Added new ssconf files with master candidate IP address
1383 (``ssconf_master_candidates_ips``), node primary IP address
1384 (``ssconf_node_primary_ips``) and node secondary IP address
1385 (``ssconf_node_secondary_ips``)
1386 - Added ``ganeti-confd`` and a client library to query the Ganeti
1387 configuration via UDP
1388 - Added ability to run hooks after cluster initialization and before
1390 - Added automatic mode for disk replace (``gnt-instance replace-disks
1392 - Added ``gnt-instance recreate-disks`` to re-create (empty) disks
1393 after catastrophic data-loss
1394 - Added ``gnt-node repair-storage`` command to repair damaged LVM volume
1396 - Added ``gnt-instance move`` command to move instances
1397 - Added ``gnt-cluster watcher`` command to control watcher
1398 - Added ``gnt-node powercycle`` command to powercycle nodes
1399 - Added new job status field ``lock_status``
1400 - Added parseable error codes to cluster verification (``gnt-cluster
1401 verify --error-codes``) and made output less verbose (use
1402 ``--verbose`` to restore previous behaviour)
1403 - Added UUIDs to the main config entities (cluster, nodes, instances)
1404 - Added support for OS variants
1405 - Added support for hashed passwords in the Ganeti remote API users file
1407 - Added option to specify maximum timeout on instance shutdown
1408 - Added ``--no-ssh-init`` option to ``gnt-cluster init``
1409 - Added new helper script to start and stop Ganeti daemons
1410 (``daemon-util``), with the intent to reduce the work necessary to
1411 adjust Ganeti for non-Debian distributions and to start/stop daemons
1413 - Added more unittests
1414 - Fixed critical bug in ganeti-masterd startup
1415 - Removed the configure-time ``kvm-migration-port`` parameter, this is
1416 now customisable at the cluster level for both the KVM and Xen
1417 hypervisors using the new ``migration_port`` parameter
1418 - Pass ``INSTANCE_REINSTALL`` variable to OS installation script when
1419 reinstalling an instance
1420 - Allowed ``@`` in tag names
1421 - Migrated to Sphinx (http://sphinx.pocoo.org/) for documentation
1422 - Many documentation updates
1423 - Distribute hypervisor files on ``gnt-cluster redist-conf``
1424 - ``gnt-instance reinstall`` can now reinstall multiple instances
1425 - Updated many command line parameters
1426 - Introduced new OS API version 15
1427 - No longer support a default hypervisor
1428 - Treat virtual LVs as inexistent
1429 - Improved job locking logic to reduce lock contention
1430 - Match instance and node names case insensitively
1431 - Reimplemented bash completion script to be more complete
1438 *(Released Thu, 4 Feb 2010)*
1440 - Fix cleaner behaviour on nodes not in a cluster (Debian bug 568105)
1441 - Fix a string formatting bug
1442 - Improve safety of the code in some error paths
1443 - Improve data validation in the master of values returned from nodes
1449 *(Released Thu, 17 Dec 2009)*
1451 - Fix security issue due to missing validation of iallocator names; this
1452 allows local and remote execution of arbitrary executables
1453 - Fix failure of gnt-node list during instance removal
1454 - Ship the RAPI documentation in the archive
1460 *(Released Wed, 30 Sep 2009)*
1462 - Fixed many wrong messages
1463 - Fixed a few bugs related to the locking library
1464 - Fixed MAC checking at instance creation time
1465 - Fixed a DRBD parsing bug related to gaps in /proc/drbd
1466 - Fixed a few issues related to signal handling in both daemons and
1468 - Fixed the example startup script provided
1469 - Fixed insserv dependencies in the example startup script (patch from
1471 - Fixed handling of drained nodes in the iallocator framework
1472 - Fixed handling of KERNEL_PATH parameter for xen-hvm (Debian bug
1474 - Fixed error related to invalid job IDs in job polling
1475 - Fixed job/opcode persistence on unclean master shutdown
1476 - Fixed handling of partial job processing after unclean master
1478 - Fixed error reporting from LUs, previously all errors were converted
1479 into execution errors
1480 - Fixed error reporting from burnin
1481 - Decreased significantly the memory usage of the job queue
1482 - Optimised slightly multi-job submission
1483 - Optimised slightly opcode loading
1484 - Backported the multi-job submit framework from the development
1485 branch; multi-instance start and stop should be faster
1486 - Added script to clean archived jobs after 21 days; this will reduce
1487 the size of the queue directory
1488 - Added some extra checks in disk size tracking
1489 - Added an example ethers hook script
1490 - Added a cluster parameter that prevents Ganeti from modifying of
1492 - Added more node information to RAPI responses
1493 - Added a ``gnt-job watch`` command that allows following the ouput of a
1495 - Added a bind-address option to ganeti-rapi
1496 - Added more checks to the configuration verify
1497 - Enhanced the burnin script such that some operations can be retried
1499 - Converted instance reinstall to multi-instance model
1505 *(Released Fri, 7 Aug 2009)*
1507 - Added ``--ignore-size`` to the ``gnt-instance activate-disks`` command
1508 to allow using the pre-2.0.2 behaviour in activation, if any existing
1509 instances have mismatched disk sizes in the configuration
1510 - Added ``gnt-cluster repair-disk-sizes`` command to check and update
1511 any configuration mismatches for disk sizes
1512 - Added ``gnt-master cluste-failover --no-voting`` to allow master
1513 failover to work on two-node clusters
1514 - Fixed the ``--net`` option of ``gnt-backup import``, which was
1516 - Fixed detection of OS script errors in ``gnt-backup export``
1517 - Fixed exit code of ``gnt-backup export``
1523 *(Released Fri, 17 Jul 2009)*
1525 - Added experimental support for stripped logical volumes; this should
1526 enhance performance but comes with a higher complexity in the block
1527 device handling; stripping is only enabled when passing
1528 ``--with-lvm-stripecount=N`` to ``configure``, but codepaths are
1529 affected even in the non-stripped mode
1530 - Improved resiliency against transient failures at the end of DRBD
1531 resyncs, and in general of DRBD resync checks
1532 - Fixed a couple of issues with exports and snapshot errors
1533 - Fixed a couple of issues in instance listing
1534 - Added display of the disk size in ``gnt-instance info``
1535 - Fixed checking for valid OSes in instance creation
1536 - Fixed handling of the "vcpus" parameter in instance listing and in
1537 general of invalid parameters
1538 - Fixed http server library, and thus RAPI, to handle invalid
1539 username/password combinations correctly; this means that now they
1540 report unauthorized for queries too, not only for modifications,
1541 allowing earlier detect of configuration problems
1542 - Added a new "role" node list field, equivalent to the master/master
1543 candidate/drained/offline flags combinations
1544 - Fixed cluster modify and changes of candidate pool size
1545 - Fixed cluster verify error messages for wrong files on regular nodes
1546 - Fixed a couple of issues with node demotion from master candidate role
1547 - Fixed node readd issues
1548 - Added non-interactive mode for ``ganeti-masterd --no-voting`` startup
1549 - Added a new ``--no-voting`` option for masterfailover to fix failover
1550 on two-nodes clusters when the former master node is unreachable
1551 - Added instance reinstall over RAPI
1557 *(Released Tue, 16 Jun 2009)*
1559 - added ``-H``/``-B`` startup parameters to ``gnt-instance``, which will
1560 allow re-adding the start in single-user option (regression from 1.2)
1561 - the watcher writes the instance status to a file, to allow monitoring
1562 to report the instance status (from the master) based on cached
1563 results of the watcher's queries; while this can get stale if the
1564 watcher is being locked due to other work on the cluster, this is
1565 still an improvement
1566 - the watcher now also restarts the node daemon and the rapi daemon if
1568 - fixed the watcher to handle full and drained queue cases
1569 - hooks export more instance data in the environment, which helps if
1570 hook scripts need to take action based on the instance's properties
1571 (no longer need to query back into ganeti)
1572 - instance failovers when the instance is stopped do not check for free
1573 RAM, so that failing over a stopped instance is possible in low memory
1575 - rapi uses queries for tags instead of jobs (for less job traffic), and
1576 for cluster tags it won't talk to masterd at all but read them from
1578 - a couple of error handling fixes in RAPI
1579 - drbd handling: improved the error handling of inconsistent disks after
1580 resync to reduce the frequency of "there are some degraded disks for
1581 this instance" messages
1582 - fixed a bug in live migration when DRBD doesn't want to reconnect (the
1583 error handling path called a wrong function name)
1589 *(Released Wed, 27 May 2009)*
1591 - no changes from rc5
1597 *(Released Wed, 20 May 2009)*
1599 - fix a couple of bugs (validation, argument checks)
1600 - fix ``gnt-cluster getmaster`` on non-master nodes (regression)
1601 - some small improvements to RAPI and IAllocator
1602 - make watcher automatically start the master daemon if down
1608 *(Released Mon, 27 Apr 2009)*
1610 - change the OS list to not require locks; this helps with big clusters
1611 - fix ``gnt-cluster verify`` and ``gnt-cluster verify-disks`` when the
1612 volume group is broken
1613 - ``gnt-instance info``, without any arguments, doesn't run for all
1614 instances anymore; either pass ``--all`` or pass the desired
1615 instances; this helps against mistakes on big clusters where listing
1616 the information for all instances takes a long time
1617 - miscellaneous doc and man pages fixes
1623 *(Released Wed, 8 Apr 2009)*
1625 - Change the internal locking model of some ``gnt-node`` commands, in
1626 order to reduce contention (and blocking of master daemon) when
1627 batching many creation/reinstall jobs
1628 - Fixes to Xen soft reboot
1629 - No longer build documentation at build time, instead distribute it in
1630 the archive, in order to reduce the need for the whole docbook/rst
1637 *(Released Fri, 27 Mar 2009)*
1639 - Now the cfgupgrade scripts works and can upgrade 1.2.7 clusters to 2.0
1640 - Fix watcher startup sequence, improves the behaviour of busy clusters
1641 - Some other fixes in ``gnt-cluster verify``, ``gnt-instance
1642 replace-disks``, ``gnt-instance add``, ``gnt-cluster queue``, KVM VNC
1643 bind address and other places
1644 - Some documentation fixes and updates
1650 *(Released Mon, 2 Mar 2009)*
1652 - More documentation updates, now all docs should be more-or-less
1654 - A couple of small fixes (mixed hypervisor clusters, offline nodes,
1656 - Added a customizable HV_KERNEL_ARGS hypervisor parameter (for Xen PVM
1658 - Fix an issue related to $libdir/run/ganeti and cluster creation
1664 *(Released Thu, 19 Feb 2009)*
1666 - Xen PVM and KVM have switched the default value for the instance root
1667 disk to the first partition on the first drive, instead of the whole
1668 drive; this means that the OS installation scripts must be changed
1670 - Man pages have been updated
1671 - RAPI has been switched by default to HTTPS, and the exported functions
1672 should all work correctly
1673 - RAPI v1 has been removed
1674 - Many improvements to the KVM hypervisor
1675 - Block device errors are now better reported
1676 - Many other bugfixes and small improvements
1682 *(Released Mon, 26 Jan 2009)*
1684 - Version 2 is a general rewrite of the code and therefore the
1685 differences are too many to list, see the design document for 2.0 in
1686 the ``doc/`` subdirectory for more details
1687 - In this beta version there is not yet a migration path from 1.2 (there
1688 will be one in the final 2.0 release)
1689 - A few significant changes are:
1691 - all commands are executed by a daemon (``ganeti-masterd``) and the
1692 various ``gnt-*`` commands are just front-ends to it
1693 - all the commands are entered into, and executed from a job queue,
1694 see the ``gnt-job(8)`` manpage
1695 - the RAPI daemon supports read-write operations, secured by basic
1696 HTTP authentication on top of HTTPS
1697 - DRBD version 0.7 support has been removed, DRBD 8 is the only
1698 supported version (when migrating from Ganeti 1.2 to 2.0, you need
1699 to migrate to DRBD 8 first while still running Ganeti 1.2)
1700 - DRBD devices are using statically allocated minor numbers, which
1701 will be assigned to existing instances during the migration process
1702 - there is support for both Xen PVM and Xen HVM instances running on
1704 - KVM virtualization is supported too
1705 - file-based storage has been implemented, which means that it is
1706 possible to run the cluster without LVM and DRBD storage, for
1707 example using a shared filesystem exported from shared storage (and
1708 still have live migration)
1714 *(Released Tue, 13 Jan 2009)*
1716 - Change the default reboot type in ``gnt-instance reboot`` to "hard"
1717 - Reuse the old instance mac address by default on instance import, if
1718 the instance name is the same.
1719 - Handle situations in which the node info rpc returns incomplete
1721 - Add checks for tcp/udp ports collisions in ``gnt-cluster verify``
1722 - Improved version of batcher:
1724 - state file support
1725 - instance mac address support
1726 - support for HVM clusters/instances
1728 - Add an option to show the number of cpu sockets and nodes in
1730 - Support OSes that handle more than one version of the OS api (but do
1731 not change the current API in any other way)
1732 - Fix ``gnt-node migrate``
1733 - ``gnt-debug`` man page
1734 - Fixes various more typos and small issues
1735 - Increase disk resync maximum speed to 60MB/s (from 30MB/s)
1741 *(Released Wed, 24 Sep 2008)*
1743 - new ``--hvm-nic-type`` and ``--hvm-disk-type`` flags to control the
1744 type of disk exported to fully virtualized instances.
1745 - provide access to the serial console of HVM instances
1746 - instance auto_balance flag, set by default. If turned off it will
1747 avoid warnings on cluster verify if there is not enough memory to fail
1748 over an instance. in the future it will prevent automatically failing
1749 it over when we will support that.
1750 - batcher tool for instance creation, see ``tools/README.batcher``
1751 - ``gnt-instance reinstall --select-os`` to interactively select a new
1752 operating system when reinstalling an instance.
1753 - when changing the memory amount on instance modify a check has been
1754 added that the instance will be able to start. also warnings are
1755 emitted if the instance will not be able to fail over, if auto_balance
1757 - documentation fixes
1758 - sync fields between ``gnt-instance list/modify/add/import``
1759 - fix a race condition in drbd when the sync speed was set after giving
1760 the device a remote peer.
1766 *(Released Tue, 22 Jul 2008)*
1768 - note: the allowed size and number of tags per object were reduced
1769 - fix a bug in ``gnt-cluster verify`` with inconsistent volume groups
1770 - fixed twisted 8.x compatibility
1771 - fixed ``gnt-instance replace-disks`` with iallocator
1772 - add TCP keepalives on twisted connections to detect restarted nodes
1773 - disk increase support, see ``gnt-instance grow-disk``
1774 - implement bulk node/instance query for RAPI
1775 - add tags in node/instance listing (optional)
1776 - experimental migration (and live migration) support, read the man page
1777 for ``gnt-instance migrate``
1778 - the ``ganeti-watcher`` logs are now timestamped, and the watcher also
1779 has some small improvements in handling its state file
1785 *(Released Fri, 13 Jun 2008)*
1787 - Experimental readonly, REST-based remote API implementation;
1788 automatically started on master node, TCP port 5080, if enabled by
1789 ``--enable-rapi`` parameter to configure script.
1790 - Instance allocator support. Add and import instance accept a
1791 ``--iallocator`` parameter, and call that instance allocator to decide
1792 which node to use for the instance. The iallocator document describes
1793 what's expected from an allocator script.
1794 - ``gnt-cluster verify`` N+1 memory redundancy checks: Unless passed the
1795 ``--no-nplus1-mem`` option ``gnt-cluster verify`` now checks that if a
1796 node is lost there is still enough memory to fail over the instances
1798 - ``gnt-cluster verify`` hooks: it is now possible to add post-hooks to
1799 ``gnt-cluster verify``, to check for site-specific compliance. All the
1800 hooks will run, and their output, if any, will be displayed. Any
1801 failing hook will make the verification return an error value.
1802 - ``gnt-cluster verify`` now checks that its peers are reachable on the
1803 primary and secondary interfaces
1804 - ``gnt-node add`` now supports the ``--readd`` option, to readd a node
1805 that is still declared as part of the cluster and has failed.
1806 - ``gnt-* list`` commands now accept a new ``-o +field`` way of
1807 specifying output fields, that just adds the chosen fields to the
1809 - ``gnt-backup`` now has a new ``remove`` command to delete an existing
1810 export from the filesystem.
1811 - New per-instance parameters hvm_acpi, hvm_pae and hvm_cdrom_image_path
1812 have been added. Using them you can enable/disable acpi and pae
1813 support, and specify a path for a cd image to be exported to the
1814 instance. These parameters as the name suggest only work on HVM
1816 - When upgrading an HVM cluster to Ganeti 1.2.4, the values for ACPI and
1817 PAE support will be set to the previously hardcoded values, but the
1818 (previously hardcoded) path to the CDROM ISO image will be unset and
1819 if required, needs to be set manually with ``gnt-instance modify``
1821 - The address to which an instance's VNC console is bound is now
1822 selectable per-instance, rather than being cluster wide. Of course
1823 this only applies to instances controlled via VNC, so currently just
1824 applies to HVM clusters.
1830 *(Released Mon, 18 Feb 2008)*
1832 - more tweaks to the disk activation code (especially helpful for DRBD)
1833 - change the default ``gnt-instance list`` output format, now there is
1834 one combined status field (see the manpage for the exact values this
1836 - some more fixes for the mac export to hooks change
1837 - make Ganeti not break with DRBD 8.2.x (which changed the version
1838 format in ``/proc/drbd``) (issue 24)
1839 - add an upgrade tool from "remote_raid1" disk template to "drbd" disk
1840 template, allowing migration from DRBD0.7+MD to DRBD8
1846 *(Released Wed, 30 Jan 2008)*
1848 - fix ``gnt-instance modify`` breakage introduced in 1.2.1 with the HVM
1850 - add command aliases infrastructure and a few aliases
1851 - allow listing of VCPUs in the ``gnt-instance list`` and improve the
1852 man pages and the ``--help`` option of ``gnt-node
1853 list``/``gnt-instance list``
1854 - fix ``gnt-backup list`` with down nodes (issue 21)
1855 - change the tools location (move from $pkgdatadir to $pkglibdir/tools)
1856 - fix the dist archive and add a check for including svn/git files in
1858 - some developer-related changes: improve the burnin and the QA suite,
1859 add an upload script for testing during development
1865 *(Released Wed, 16 Jan 2008)*
1867 - experimental HVM support, read the install document, section
1868 "Initializing the cluster"
1869 - allow for the PVM hypervisor per-instance kernel and initrd paths
1870 - add a new command ``gnt-cluster verify-disks`` which uses a new
1871 algorithm to improve the reconnection of the DRBD pairs if the device
1872 on the secondary node has gone away
1873 - make logical volume code auto-activate LVs at disk activation time
1874 - slightly improve the speed of activating disks
1875 - allow specification of the MAC address at instance creation time, and
1876 changing it later via ``gnt-instance modify``
1877 - fix handling of external commands that generate lots of output on
1879 - update documentation with regard to minimum version of DRBD8 supported
1885 *(Released Tue, 4 Dec 2007)*
1887 - Log the ``xm create`` output to the node daemon log on failure (to
1888 help diagnosing the error)
1889 - In debug mode, log all external commands output if failed to the logs
1890 - Change parsing of lvm commands to ignore stderr
1896 *(Released Wed, 28 Nov 2007)*
1898 - Another round of updates to the DRBD 8 code to deal with more failures
1899 in the replace secondary node operation
1900 - Some more logging of failures in disk operations (lvm, drbd)
1901 - A few documentation updates
1908 *(Released Tue, 13 Nov 2007)*
1910 - Change configuration file format from Python's Pickle to JSON.
1911 Upgrading is possible using the cfgupgrade utility.
1912 - Add support for DRBD 8.0 (new disk template ``drbd``) which allows for
1913 faster replace disks and is more stable (DRBD 8 has many improvements
1914 compared to DRBD 0.7)
1915 - Added command line tags support (see man pages for ``gnt-instance``,
1916 ``gnt-node``, ``gnt-cluster``)
1917 - Added instance rename support
1918 - Added multi-instance startup/shutdown
1919 - Added cluster rename support
1920 - Added ``gnt-node evacuate`` to simplify some node operations
1921 - Added instance reboot operation that can speedup reboot as compared to
1923 - Soften the requirement that hostnames are in FQDN format
1924 - The ``ganeti-watcher`` now activates drbd pairs after secondary node
1926 - Removed dependency on debian's patched fping that uses the
1927 non-standard ``-S`` option
1928 - Now the OS definitions are searched for in multiple, configurable
1929 paths (easier for distros to package)
1930 - Some changes to the hooks infrastructure (especially the new
1931 post-configuration update hook)
1932 - Other small bugfixes
1934 .. vim: set textwidth=72 syntax=rst :